[tracker1]

IMO, the greatest force multiplier in onboarding is automation... How many user interactive steps does a dev need to walk through to get up and running?

I've often gotten this down to a single directory and a primary script you need to run 3-4 times... mostly because of required reboots, such as after WSL setup, also, a few options to set in Docker Desktop. I've done similar with mac/homebrew.

From there, you have docker-compose and shell scripts to run against the projects... these scripts and the compose file(s) are the entry points into a project, actually documenting where things are, and how they connect (to an extent).

Effectively, getting a working solution running sooner than later. This is just my own opinion, but automating a dev environment for onboarding, especially if you're a larger org, is a great tactical decision.

[xp84]

Heavily seconding the “setup script” method for anything that can be scripted in lieu of documentation and handholding. I’ve done this at multiple companies and it proved valuable. Even if it eventually breaks it’s easy to fix, and it’s usually pretty self-documenting, so anyone who’s even a little curious what’s going on, they have access to all the “secrets” of how it’s being done (and can also improve it).